Study identifies optimal C2 recovery for NGL plants

The feed gases enter the plant from the surface production facilities that process the gas to various degrees. Depending on the gas processing network and distances to the field, the feed gas may already be dehydrated and possibly hydrocarbon dewpointed. In other cases, the feed gas may be wet with possibly two or three phases.

Downstream of the slug catchers or inlet separators, feed gas is typically filtered and acid gases are removed through amine processing. The acid gas is then further processed in a sulfur-recovery unit.

After sweetening, the process gas may be compressed and cooled prior to dehydration and mercury removal. The degree of compression, type of dehydration, and need for mercury removal all depends on the type of NGL-recovery process used.

Expander based C2-recovery processes usually require dehydration to less than 5-ppm water (sometimes less than 1 ppm) to avoid water freezing in low-temperature process equipment. This usually requires molecular sieve dehydration with high-temperature regeneration.

Ethane-recovery plants usually use aluminum plate-fin heat exchangers because they offer high performance for a relatively low cost. Mercury must be removed to a concentration of 10 nanograms/cu m to avoid damaging aluminum process equipment.

The pretreated process gas is now ready to feed the NGL-extraction facility.

NGL-extraction technologies

Three classes of processes for deep recovery of NGL include refrigeration, expander-based cryogenic, and lean-oil absorption. Expander-based plants tend to dominate C2+ NGL recovery because they can achieve deep recoveries at lower capital and operating expense.

Mechanical refrigeration and lean-oil processes are more competitive when feed gases are rich and available at moderate pressure, typical of those in the Middle East and Nigeria.

Basic refrigeration processes are flexible and appropriate for various levels of NGL recovery-from simple dewpointing to deep C2 recovery depending on the refrigeration cycle.1

Multi-stage C3 chilling is typically used for C3+ recoveries. C2+ recovery generally requires lower refrigeration temperatures with mixed refrigerant or cascade cycle processes. These processes, however, are usually not competitive with expander-based processes except for rich or moderate-pressure feeds (20-30 bar).

Turboexpander-based processes rely on isentropic expansion of the precooled feed gas to generate the cold temperatures required to effect NGL separation and recovery.1-3 These processes dominate NGL recovery, particularly when C2 or deep C3 recovery is required. Expander processes are generally favored for leaner, higher-pressure feed gases.

Refrigerated lean-oil absorption processes use a refrigerated liquid stream of heavier hydrocarbons (lean oil) to absorb the C2+ or C3+ components from a precooled feed stream in an absorber (OGJ, Sept. 29, 1997, p. 86; Mar. 1, 1999, p. 62; Oct. 29, 2001, p. 56).4

The bottoms product from the absorber is regenerated in a second column where the NGL product stream is the overhead product and some of the bottoms product is used as the regenerated lean oil stream after it is cooled and pumped.

The relative merits of refrigerated lean oil processes generally improve for rich feed gases available at pressures less than 31 barg (450 psig). Although lean-oil absorption is rarely, if ever, selected instead of expander processes for C2 recovery, there is continued interest and development in this technology.

C2-recovery parameters

The appropriate C2-recovery level is normally an economic decision that is a function of many variables including the regional value of C2, the richness of the feed, existing processing facilities, the type of NGL recovery plant used, and the plant owner’s own experience.

Regional value of C2

The regional value of C2 relative to sales gas is the economic basis for its recovery. It is often difficult to establish a value for C2, however, because it depends on demand from the petrochemical industry.

One must ask what is the existing commercial value for C2 relative to the value if retained in the sales gas. Sales gas is often sold on a calorific value, in which case the differential value of a pure C2 product is relative to its inclusion in a sales-gas stream.

In regions with well-established petrochemical industries, the C2 value fluctuates with ethylene demand. In areas with developing petrochemical industries, C2 value may change dramatically when a steam cracker is built.

The presence of steam crackers and a petrochemical infrastructure usually guides the market conditions for ethane. This creates a “chicken and egg” dilemma in which the petrochemical industry depends on the availability of ethane and the economic viability of C2 recovery is based on a C2 market that does not yet exist. In emerging petrochemical regions, a C2-recovery facility may initially be justified on the basis of the future value for C2.

Deeper C2 recovery generally comes at increased initial capital and operating costs for a given process. Ultimately, higher C2 recoveries are justified on a net present value or internal rate of return basis for higher C2 values relative to sales gas.

Feed-gas conditions

Percent C2 recovery is commonly used to describe a NGL recovery plant’s performance; but it can be misleading. For many projects there is a tendency to set a C2-recovery level early on the basis of perceived technical feasibility and convenience.

This approach can result in a suboptimal conceptual design; some designs lose product that could be recovered readily at minimal incremental cost. In other designs, excessive capital and operating costs result from a specification that seemed reasonable initially but was actually too high.

Feed conditions, particularly the feed richness, are critical in determining the C2-recovery level. Richer gas offers the potential to extract more liquid, which can lead to higher NGL product revenues. This increased liquid recovery also requires larger equipment and more power to process to the same recovery level.

Variable richness in the feed gas is one of the main reasons for the inappropriate application of C2-recovery percentages. In general, a process can typically achieve a higher percent recovery for a richer feed. This is not always the case, however; rich feeds create processing challenges associated with high liquid fractions in the cold separator.

We recently designed, for example, two open-art NGL recovery processes for different projects. In one case, C2 recovery was specified at 88%. A similar process specified 95% recovery for a richer feed; a seemingly more aggressive specification.

An examination of the separation, however, showed that the plant with 88% recovery actually accomplished a more challenging separation based on separation factor and relative volatility of the key (C1:C2) components.

Fit into overall operations

The NGL-recovery process as it integrates with the overall process is another consideration. In many cases, the recovery process is defined during the design’s early phases. In other cases, client preference for an open-art process or a particular licensor or process may constrain process selection.

The plant’s fit into overall operations is also a key consideration. Important questions that one should evaluate and answer in the early design phases include: What types of plants are currently operating in the region? What operational experience does the owner have? Is there an opportunity to debottleneck or retrofit an existing plant or is a new plant required? Does the plant require a C2-rejection mode of operation?

C2-recovery step changes

There are several incremental increases (step changes) in the cost of NGL recovery. For hydrocarbon dewpointing and shallow NGL recovery these are typically the limit to which gas can be dewpointed using auto-refrigeration with a Joule-Thomson expansion and refrigeration system.

For deeper NGL recovery, other options continue to have major effects on the economic viability of the process. These step changes relevant to C2 recovery include the NGL-recovery process chosen, CO2 tolerance and rejection, refrigeration system configuration, and materials of construction.

NGL-recovery process

The NGL-recovery scheme selected is likely to be the cause of the most important step change increase in the costs associated with C2-recovery levels. In some cases, older open-art processes are suitable for a particular application. The conventional expander process and gas subcooled processes are both open art and capable of moderate to deep C2 recovery.

The results of this study are based on open-art NGL extraction processes. In many cases, licensed processes that are protected by patents offer superior economic results, particularly for deep C2 recovery. Residue-gas recycle processes are proprietary and offer several advantages for deep C2 recovery including the potential for less power consumption, lower capital cost, and increased CO2 tolerance.

CO2 tolerance, rejection

Product specifications or the amount of CO2 freeze-out in the process may constrain the degree to which CO2 must be removed from the feed gas. The general rule of thumb is that CO2 travels with the C2 throughout the process.

When a C2 product is generated downstream, the CO2 specification in the C2 usually determines the CO2 that can slip through the process. When C2 is rejected to sales gas, the calorific value or possibly a CO2 specification in the sales gas determines the level of CO2 rejection.

For deep C2 recovery, lower operating temperatures are required when expander-based processes are used. The triple point for CO2 is about -77° F.; operating below this temperature may cause solid CO2 to collect in the system. CO2 deposition from the liquid (or gas) phase can allow CO2 to build up and foul heat exchangers or columns and increase pressure drop through process equipment.

The designer should consider the use of a selective amine to slip some CO2 into the NGL-recovery plant’s feed gas. This may decrease the cost of the acid-gas removal equipment or eliminate the need for co-firing or oxygen enrichment in the sulfur-recovery unit.

Refrigeration system

For expander-based processes, supplemental refrigeration represents a large step change in the plant’s capital and operating costs. In some cases when the feed gas is available at relatively high pressures and is relatively lean, supplemental refrigeration may not be required even for relatively deep C2 recovery. When richer feed gases are processed, extensive supplemental refrigeration is often required regardless of the recovery target.

If supplemental refrigeration is used, one must consider the complexity of the refrigeration system. Multiple-pressure-level refrigeration systems are more efficient but are generally more capital intensive and complex.

Refrigeration systems with more stages have higher efficiencies, which may justify a higher capital cost. Often, one must choose between two and three-stage refrigeration.

The number of refrigeration stages depends on the cooling-load temperatures and overall refrigeration requirements.

Materials of construction

Metallurgical improvements can result in distinct cost increases for individual equipment items.

For the NGL-recovery process, the feed gas has been sweetened, dehydrated, and filtered; therefore, the metallurgical considerations are primarily based on design temperatures. Design temperatures for process equipment are a function of operating temperatures and the type of service.

In general, deeper C2 recoveries require lower process temperatures and therefore more expensive materials to retain ductility. Aluminum is suitable for low temperature, plate-fin exchangers but can suffer from drawbacks associated with a high coefficient of thermal contraction and low mercury tolerance.

C2-recovery case study

Our C2-recovery level study was for a world-scale gas plant as part of an overall conceptual design. The NGL-recovery plant processed a blend of associated gas streams and overheads from a sour-condensate stripper. After preliminary process selection was complete, the client asked us to consider the appropriate C2-recovery level for open art, expander based, NGL-recovery processes.

The objective was to generate data for incremental capital costs as a function of C2 recovery. We could then consider these incremental capital costs in conjunction with operating costs and the process impact on other revenue streams (C3+ recovery); this would support our target for the C2-recovery level for the plant based on economic modeling.

Because we conducted the evaluation for preselected processes it illustrates clearly the cost impact of C2-recovery level.

Study scope, methodology

The scope was limited to subsystems that were sensitive to C2-recovery level. Fig. 2 shows the study area.

The facilities considered included these subsystems:

• The expander-based NGL plant including the compressor-loaded expander, aluminum heat exchangers, columns, cold separator, and associated process equipment.

• The refrigeration system that provides cooling to the NGL plant to support NGL recovery and to the pretreatment facilities for chilling upstream of the dehydration beds.

• The sales-gas compression system including the suction scrubbers, compressors, and aftercoolers required to provide sales gas to the grid at 68 barg (1,000 psig) and meet any plant residue recycle requirements.

• The NGL storage and export facilities including NGL storage, NGL transfer pumps, and NGL export pumps.

Although deep C2 recovery was specified for the project, the client was interested in assessing C2 recovery for the widest possible range. This evaluation required us to include two process schemes-a typical two column, C3+ NGL-recovery process and a single-column process typical of those used for moderately deep C2 recoveries.

We then modified these simulations either to increase or decrease the C2-recovery level to provide coverage for the entire possible range. From the two base simulations, we considered a total of 12 cases at C2-recovery levels of 0.3-99.3%.

Development of sized equipment lists and estimates for all 12 cases would have been too time consuming and expensive. To expedite the study for the time-critical project, we needed reliable cost estimates without a full parametric analysis. We therefore used several methods including the development of cost correlations based on key simulation outputs.

The key measurements were extracted from the simulations and used to develop cost estimates using the correlations. Key information relating to plant costs was extracted from the simulations including variables such as gas and liquid capacities, refrigeration compressor horsepower, and C2-recovery level.

These data were then converted into differential costs using a correlation-based costing model. Costs were then plotted against C2-recovery level for the two processes. Other important factors such as C3 recovery and plant compression power were also extracted from the simulations.

C2 process simulation

The base simulation was developed and optimized at 95% C2 recovery in a C2+ NGL product stream. We ran eight cases using this base simulation with minor process changes. These cases ranged from 50% to 99+% C2 recovery.

For each case, we optimized the model using the same process conditions and constraints. We changed C2 recovery levels by adjusting variables such as the cold separator temperature. Process conditions such as the refrigerant circulation rate and exchanger terminal temperatures were adjusted to satisfy simulation constraints such as exchanger temperature approaches.

C3 process simulation

The expander-based C3+ recovery process was developed and originally optimized for ~97% C3 recovery in a C3+ NGL product stream containing minimal C2.

We ran two additional cases using this basic simulation with relatively minor process changes to allow recovery up to about 75% of C2. We changed the C2-recovery levels by adjusting the reboiler product ratio specification and making other process changes while satisfying process constraints.

Findings

The costs were expressed relative to the C3+ recovery-only case as a function of C2 recovery for the two processes (Fig. 3). Measuring costs relative to no C2 recovery allows one to evaluate the cost associated with C2 recovery and does not imply a base case at low C3 recovery.

Capital costs decreased at lower recoveries and rapidly increased once a process was pushed beyond a range for which it was well suited. The process scheme initially developed for C3+ recovery shows a large incremental C2 recovery with minimal increased capital cost when the column bottoms specification was changed to recovery rather than C2 rejection.

The process also rapidly increases in cost as the C2-recovery level rises beyond simple incremental values. This confirms our expectation that the C3+ recovery process is poorly suited for deeper C2 recovery.

The process scheme initially developed for C2+ shows the same rapid increase in capital costs as C2 recovery increases beyond about 96%. At first glance, the rapidly decreasing C2-recovery capital cost for the process at lower recovery levels is unexpected. When considered in conjunction with the lower C3 and C4 recovery for the process, however, the result makes more sense.

The C2+ process relies on a single column for NGL separation. The phase equilibria in the top stage therefore determined the C3+ carryover into the sales gas. This limitation means that the C2+ recovery process is not suitable for C2 recoveries less than about 65%.

As C2 recovery decreases for the C2+ process (Fig. 3), C3+ recoveries also decrease (50% C2 recovery results in 86% C3 recovery).

The general shape of the C2+ curve suggests that C2 recovery up to about 96% may be reasonable from a capital cost perspective. The roughly constant slope of the curve between 65% and 96% suggests that there is no compelling reason to a design that recovers less than 96% of the C2 (i.e. if it is economical to recover 75% of the C2, then it is also economical to recover it at a 95% level).

These results are supported by a similar trend for plant horsepower, which was used as an indication of operating costs. Ultimately, operating and capital costs and product values were used to develop an economic basis for C2-recovery level.

Representing costs as a continuous function is incorrect in the strictest sense because step-changes in costs mean that there are discontinuities that are ignored. Representation as a curve gives an overall impression of the costs trends but the area between simulated points are interpolated.

The study also considered the C3 and C4-recovery level (product revenues) and power consumption (as an indication of operating cost). As C2 recovery decreases in the C2-recovery process, C3+ recovery also decreased because the phase equilibria in the top stage of the column ultimately determine the C3+ carryover into the sales gas.

Process-gas compression power, refrigeration power, and other key factors related to operating costs were also taken from the simulation. The results showed that the operating costs varied about 35%; deep C2 recovery required the most energy.

The C2-recovery process used to recover about 50% of the C2 had the lowest operating costs but also had relatively low recovery of C3+ components. ✦

References

1. “GPSA Electronic Data Book,” Vol. II, Section 16-Hydrocarbon Recovery, 11th Ed., Tulsa: Gas Processors Association, 1998.

2. Wilkenson, J.D., Hudson, H.M., Cuellar, K.T., and Pitman, R.N., “Next Generation Processes for NGL/LPG Recovery,” presented to the 77th Annual Gas Processors Association Conference, Mar. 16-18, 1998, Dallas.

3. Wilkenson, J.D., “NGL Recovery and Gas Processing,” presented to the 8th Doha Technical Symposium, November 1984, Doha, Qatar.

4. AET Inc. list of publications, http://www.aet.com/papers.htm, Oct. 3, 2004.

The author

Michael Barclay ([email protected]) is a senior process engineer in the oil and gas division for Foster Wheeler Ltd., Reading, UK. He has 10 years’ experience in gas plant design including NGL recovery, natural gas liquefaction, and sour-gas treatment. He holds a BS from the University of Wisconsin, Madison, and an MS from Pennsylvania State University. He is a senior member of the AIChE, and a member of the ASME and the Cryogenic Society of America.